LSAY work and school - ERIC · 2013. 8. 2. · 29 ANOVA for TER scores against Year 12 working hours, males 47 30 Regression means for TER against Year 12 working hours, ... Chang

Does combining school and work affect school and post-school outcomes? ALISON ANLEZARK

PATRICK LIM

LONGITUDINAL SURVEYS OF AUSTRALIAN YOUTH

NCVER

Does combining school and work affect school and

post-school outcomes?

Alison Anlezark Patrick Lim

National Centre for Vocational Education Research

Funded by the Australian Government Department of Education, Employment and Workplace Relations with support from state and territory governments.

The views and opinions expressed in this document are those of the author and do not necessarily reflect the views of the Australian Government or

state and territory governments.

.

Publisher’s note Further information regarding the Longitudinal Surveys of Australian Youth (LSAY) can be found at

.

© Commonwealth of Australia, 2011

This work has been produced by the National Centre for Vocational Education Research (NCVER)

on behalf of the Australian Government and state and territory governments with funding provided

through the Australian Department of Education, Employment and Workplace Relations. Apart from

any use permitted under the Copyright Act 1968, no part of this publication may be reproduced by any

process without written permission of the Commonwealth. Requests should be made to NCVER.

The views and opinions expressed in this document are those of the author(s) and do not necessarily

reflect the views of the Australian Government or state and territory governments.

ISBN 978 1 921809 12 9 web edition

ISBN 978 1 921809 18 1 print edition

TD/TNC 104.12

Published by NCVER

ABN 87 007 967 311

Level 11, 33 King William Street, Adelaide SA 5000

PO Box 8288 Station Arcade, Adelaide SA 5000, Australia

ph +61 8 8230 8400 fax +61 8 8212 3436

email [email protected]

NCVER

About the research

Does combining school and work affect school and post-school outcomes? Alison Anlezark and Patrick Lim, NCVER

One of the distinctive characteristics of Australia’s secondary schooling system is the sizable proportion of students working part-time. This phenomenon raises important policy issues: does working part-time assist or hinder academic performance? Does it assist the transition to the labour market? This report uses data from the Longitudinal Surveys of Australian Youth (LSAY) of students who were aged 15 in 2003 to look at these questions.

Key messages: Students who are combining work and school, on average, work 11–12 hours a week, with more

females working than males; however, on average, males who are combining work and school work longer hours.

Combining school and work has a modest negative impact on school and post-school study outcomes when hours are long (in excess of 15–20 hours a week). Females are better able to balance school and work, with the magnitude of these negative effects generally being less than for males.

Working for relatively few hours a week (around five hours per week) has a positive impact on post-school full-time employment, compared with not working at all. Females have to work slightly longer hours to realise maximum benefits from working (15–20 hours per week) than males (10–15 hours per week), but the magnitude of the effect is comparable with males.

While one has to be cautious in attributing causation, it does appear that students who are working lengthy hours in part-time employment are signalling an orientation towards employment and away from formal education.

Tom Karmel Managing Director, NCVER

NCVER 5

Contents

Tables and figures 6 Introduction 8

Background 8

Research approach 11 Statistical approach 12

How many students are working? 14 Distribution of hours worked 15

Who combines part-time work and school? 18 Characteristics of students who combine school and work and propensity score regression 19 Average hours worked by student characteristics 20

School outcomes 22 Impact on school retention 22 Impact on school performance (TER score) 23 Summary 24

Post-school outcomes 25 Effect of working in Year 12 on post-school full-time study status 25 Effect of working in Year 12 on post-school full-time employment status 26 Summary 27

Discussion 28 References 29 Appendices

A 30 B 32 C 50

6 Does combining school and work affect school and post-school outcomes?

Tables and figures

1 Tables

Percentage of respondents working in each school year level, Y03, 2003–07 15

2 Summary statistics of working and working hours by year level, Y03 cohort 16

3 Summary of characteristics of those who combine school and work from previous research 18

4 Average hours worked in Year 12 by student characteristics, by gender 21

5 Predicted probability of Year 10 to Year 11 retention by hours worked in Year 10 22

6 Predicted probability of Year 11 to Year 12 retention by hours worked in Year 11 23

7 Mean TER scores by hours worked in Year 12, males 238 Mean TER scores by hours worked in Year 12, females 249 Predicted probability of undertaking full-time post-school study

for hours worked in Year 12, males 2510 Predicted probability of undertaking full-time post-school study

for hours worked in Year 12, females 2611 Predicted probability of full-time employment with no post-

school study for hours worked in Year 12, males 2612 Predicted probability of full-time employment with no post-

school study for hours worked in Year 12, females 2713 LSAY Y03 data by school year level and year of data collection

(weighted) 3014 Summary statistics of working hours and numbers by year level

by gender 3115 Regression results for working in Year 10: males, Y03, 2003–07 3316 Regression results for working in Year 10: females, Y03,

2003–07 3517 Regression results for working in Year 11: males, Y03, 2003–07 3718 Regression results for working in Year 11: females, Y03,

2003–07 3919 Regression results for working in Year 12: males, Y03, 2003–07 4120 Regression results for working in Year 12: females, Y03,

2003–07 4321 Type 3 analysis of effects for Year 11 retention, males 4522 Regression results Year 11 retention, males 4523 Type 3 analysis of effects for Year 11 retention, females 4524 Regression results Year 11 retention, females 4525 Type 3 analysis of effects for Year 12 retention, males 4626 Regression results Year 12 retention, males 4627 Type 3 analysis of effects for Year 12 retention, females 4628 Regression results Year 12 retention, females 46

NCVER 7

29 ANOVA for TER scores against Year 12 working hours, males 47

30 Regression means for TER against Year 12 working hours, males 47

31 ANOVA for TER scores against Year 12 working hours, females 47

32 Regression means for TER against Year 12 working hours, females 47

33 Type 3 analysis of effects for post-school study, males 4834 Regression results full-time study post-Year 12, males 4835 Type 3 Analysis of effects for post-school study, females 4836 Regression results full-time study post-Year 12, females 4837 Type 3 analysis of effects for labour market outcomes: no

full-time study post-Year 12 for Year 12 working hours, males 4938 Regression results full-time employment post-Year 12, males 4939 Type 3 analysis of effects for labour market outcomes: no

full-time study post-Year 12 for Year 12 working hours, females 4940 Regression results full-time employment post-Year 12, females 4941 Year 12 completion status by intensity (hours) worked per

week in Year 10, Y03 cohort in 2007, males 5042 Year 12 completion status by intensity (hours) worked per

week in Year 10, Y03 cohort in 2007, females 5143 Hours of work in Year 12 and later labour market outcomes

for Y03 in 2007: no post-school study, males 5144 Hours of work in Year 12 and later labour market outcomes

for Y03 in 2007: no post-school study, females 52

1 Figures

Proportion of 15 to 19-year-olds at school who are employed, August 1986–2008 14

2 Box plot of working hours for all respondents by school year level, Y03 cohort 16

3 Box plot for working hours by school year level by gender, Y03 cohort 31


Introduction Background The proportion of young people combining school and work is on the increase. Depending on the sources consulted (for example, ABS labour force statistics or LSAY Y03 cohort), the actual proportions of school students working are estimated as at between 30% and 60%.

The increase in young people combining school and work can be explained on the demand side by the changing structure of the Australian workforce, with employers seeking more flexible, casual workers, particularly in the hospitality and retail sectors, for which young people are well suited (Biddle 2007).

On the supply side, there is a plentiful supply of young people who are staying on at school and who see part-time work as a means of gaining some financial independence from their parents. Young people make decisions on whether or not to work, based on the availability of work, their desire for financial independence, their ability to travel to the work location and whether or not their parents want them to work. Rarely are the jobs young people choose to work in while at school selected as intentional career pathways (Robinson 1999; Smith & Green 2005; Howieson, McKechnie & Semple 2006).

We see there is a good match between supply (young people) and demand (employers) for student workers, but is this a good thing for young people? Are students able to manage the competing demands of combining school and work? Does combining school and work have a beneficial or detrimental impact on their school and post-school outcomes? The purpose of this paper is to explore these questions.

The majority of previous Australian (see, for example, Robinson 1996, 1999; Vickers, Lamb & Hinkley 2003) and international research (Howieson, McKechnie & Semple 2006; Marsh & Kleitman 2005; Singh, Chang & Dika 2007) finds that combining school and work has a negative impact on school performance. In general, the more hours worked, the more negative the effect.

It is not difficult to understand why combining school and work can be detrimental to school performance. Hours of study are foregone by working, and students may be distracted by work, or too tired to concentrate properly at school. However, Marsh and Kleitman (2005) take this one step further and suggest that it may also be what young people do with the money they earn that can be detrimental: the researchers find that access to money can lead to an increase in anti-social behaviour such as drug taking and alcohol abuse, which in turn can affect school performance.

But how much is too much work? The Australian studies cited above use different longitudinal datasets to analyse the effects of working in specific (but different) school year levels. These approaches make cross-study comparisons difficult, and it would be naive to assume that the characteristics of students who work in Years 9 or 10 are the same as those who work in Years 11 or 12, and that their effects will be the same. Vickers, Lamb and Hinkley (2003)1

1 Using the LSAY Y95 cohort data, an aged-based cohort who were aged 15 years in 1995.

found that working more than five hours a week in Year 9 had a detrimental effect on Year 12 completion.

NCVER 9

Robinson (1999)2

A study conducted in the United States by Marsh and Kleitman (2005)

found working more than ten hours a week in Years 11 and 12 negatively affected tertiary entry rank (TER) scores, and working in Year 11 for more than ten hours a week affected Year 12 completion.

3

Staff and Mortimer (2007)

found that the number of hours worked per week may be as high as 20 before the negative effects of combining school and work are felt. In this study, students who worked at least 20 hours a week in high school reported fewer hours of homework and lower test scores than students who limited their hours.

4

In addition to hours of work, it is also important to understand the characteristics of those who work when at school. If we find that working when at school is beneficial, then we might want to promote combining school and work to those groups of young people not currently engaged in this activity. Similarly, if we find that combining school and work is detrimental, then we may want to identify those at risk from this activity. It is also important to consider the intensity of work while at school. That is, do young people who work longer hours when at school have different characteristics from those who work fewer hours, or not at all? Are some people able to tolerate work when at school more than others? Research by Shanahan and Flaherty (2001) found that a well-rounded youth often combines some paid work with school extracurricular activities, with no negative effect on school performance.

suggest that it may not just be the number of hours, but the intensity and duration of the work, classified as ‘occasional’, ‘sporadic’, ‘steady’ and ‘most invested’, which affects school performance. This is one of the few studies that finds that combining school and work can improve school outcomes. Their US research reports that part-time work during high school can set good patterns of work–study combinations, and moderate but steady combinations of school and work can facilitate educational attainment for some underperformers.

Most of the benefits arising from combining school and work appear to be for post-school employment rather than any school-related benefits. Marsh and Kleitman (2005) find a reduction in post-secondary unemployment for students who combine school and work. Billett (2006) finds that part-time paid work and school can teach students about the world of work and broaden their understanding of post-school options and pathways, but the types of work seem to matter. Most young people work predominantly in the hospitality (fast food) and retail sectors, which may allow for the development of some ‘employability’ skills (Biddle 2007). However, many of these jobs require young people to work predominantly with their peers, and there is little evidence that combining school with these types of jobs prepares young people for the world of work (Meyerhoff 2006).5

Finally, the advantages and disadvantages of combining school and work are not clear-cut. Some young people are better able to manage the competing demands of combining school and work. Young people from more advantaged backgrounds, who are in general more strongly focused on an academic trajectory, are more likely than their less advantaged counterparts to work, but work for fewer hours, and generally have more positive school and post-school study outcomes anyway. In contrast, youth from more disadvantaged backgrounds and those with poorer grades and lower educational aspirations are more likely to work longer hours when at school and have poorer school and post-school outcomes (Staff & Mortimer 2008). However, we do not know whether their choice to work longer hours is influenced by their poor school performance, or whether their

2 Using the Youth in Transition 1975 birth cohort in 1994 at age 19 years. 3 Using the US National Education Longitudinal Survey (NELS) of a 1988 cohort. 4 Using the US Youth Development Study, a longitudinal survey of 1010 grade 9 students and parents from Minnesota,

from 1987, followed from age 19 to 31 years. 5 The topic of combining school and work is very broad; an area not explored in this paper, but worthy of future

research could be to consider good work versus bad work by considering employment type and when the employment occurs.


poor school performance is the result of longer hours worked. That is, are these individuals already disengaged from schooling, so that, essentially, working does not affect their school outcomes?

In this paper we seek to update the existing research on the impact of combining school and work. Our analysis is disaggregated by gender and provides a more nuanced measure of hours worked and its relationship to outcomes than previous research by looking at the effect at each school year level.

In the first part of the paper we describe the statistical approach, and then quantify and describe the distribution of hours worked when students are at school between Years 9 and 12. This shows how many students are combining school and work and provides an understanding of how hours of work change between the school year levels. We then provide a summary of the characteristics of students who combine school and work to complete the picture.

In the main part of the paper we look at the effect of hours of work on school and post-school outcomes, allowing for the background and aspirational characteristics of the individual. The effect of combining school and work on retention to Years 11 and 12, as well as TER scores, is analysed to measure effects on school outcomes. Post-school outcomes are measured in terms of full-time post-school study and full-time employment for Year 12 completers. We conclude the paper with a discussion on what new evidence this paper brings to the debate on part-time work and school.

Consistent with previous research, we find some negative effects from combining school and work on school and post-school study outcomes for those working longer hours, but positive effects on post-school employment.

NCVER 11

Research approach In updating the previous research, we build on the earlier work of Vickers, Lamb and Hinkley (2003) and Robinson (1999) using longitudinal datasets, but we use a more recent cohort of young people,6

However, comparing the results with the previous research is difficult because different approaches and cohorts were used:

the Y03 cohort, which is a group of 10 370 young people who were aged around 15 in 2003.

Vickers, Lamb and Hinkley (2003) used data from the LSAY Y95 cohort, focusing on Year 9 students (the first wave of the cohort) and included the whole cohort and did not exclude early school leavers. Males and females were modelled separately. They did not report on the effects of students working in different school year levels, but focused only on combining school and work in Year 9, and the effect this had on Year 12 completion and post-school employment outcomes in the first few years beyond school. The second part of their study focused on the effect of work on post-school university students.

Robinson (1999) used data from the Youth in Transition 1975 birth cohort, but took a broader approach and first looked at motivations for working.7

Both of these previous studies described the characteristics of those who combined school and work, and then controlled for these characteristics in their models.

She then measured the effects of combining school and work in Years 11 and 12 on Year 12 completion and Year 12 results (TER scores), but did not run separate models for males and females. She then went on to look at the effect on post-school outcomes, measured as incidences of unemployment and income and job type.

In this paper we conduct some analysis which has not been previously undertaken, by looking at the effect of working on retention to Years 11 and 12 by gender.8

Post-school outcomes are analysed in a different way from Robinson (1999), by restricting our analysis to Year 12 completers and testing whether working at school in Year 12 has a positive or negative effect on the likelihood of going on to post-school full-time study or full-time employment. This approach provides a more direct relationship between the year the student combined school and work and the post-school outcomes, and explores whether working is beneficial for those not pursuing an academic trajectory after Year 12. Separate models are again run for gender.

We then model the effects of working in Year 12 on TER score in a similar manner to the work of Robinson (1999), but with separate models for males and females.

We do not consider the effects of working in Year 9 on either school or post-school outcomes (as did Vickers, Lamb & Hinkley 2003), because the Y03 LSAY dataset is an age-based rather than a year-based cohort (as was the Y95 cohort), and there are too few students to analyse in the Y03 cohort in

6 Refer to appendix A for detail on the data and scope. 7 This information was captured from a series of questions asked of the YIT cohort in 1992 (when they were aged 17

years) about their experiences of being a part-time worker. The questions were phrased as a series of statements commencing with I work because … These questions are not asked of the current LSAY cohorts (Y95, Y98, Y03, and Y06).

8 We did consider looking at Year 12 completion, but because the majority (98%) of LSAY students who commenced Year 12 completed it, there was little scope for work to affect Year 12 completion.


Year 9. However, we do analyse by school year level rather than age, for consistency with the Vickers, Lamb and Hinkley (2003) study.

Our choice of a measure of work is selected as a range of hours worked in any given year level. This approach provides greater sensitivity than using a single measure of cumulative hours worked across all year levels, or a binary variable of work and no work. Hours of work are summed across all jobs, and because the LSAY interviews are conducted between July and January each year, they may also include school holiday jobs. However, due to the timing of the LSAY interviews, only the September school holidays would be captured for the majority of respondents. We therefore considered it important to include work during this time because this is when most senior school students are preparing for end-of-year exams.

Statistical approach In this paper we use a series of gender-specific regression models to describe the characteristics of those who are most likely to combine school and work in each school year level between Years 10, 11 and 12. From these models we derive propensity scores to control for background characteristics in the later models of post-school outcomes.

The approach taken in this work is to use a methodology that treats the hours worked in each year level as random treatments. Unfortunately, LSAY is not a traditional experimental design in which each treatment level is randomly assigned to experimental units (individuals). The aim of randomisation is to ensure that any pre-existing background effects (such as achievement, socioeconomic status etc.) are assigned evenly across each of the treatment levels. That is, randomisation would ensure that those who are working in Year 10 are not all from a single socioeconomic status or achievement level. The primary way of achieving this balance in an observational study is to use propensity score weighting (Rosenbaum & Ruben 1983). The propensity scores are fitted as covariates in regression analyses to ensure that the background of individuals is ‘balanced’ across the treatment groups of interest.

Propensity scores are assigned to each individual in the cohort, where the propensity score is the inverse of the probability of working in the relevant school year level (probability of not working could also be used). For this study, propensity scores were derived for working in each of Years 10, 11, and 12 separately for males and females. A series of logistic regressions, in which the response variable is working or not working, were fitted against the following background characteristics:

school sector

locality

socioeconomic status (parental occupation)

academic achievement (in maths, problem-solving, reading and science)

participation in VET in Schools in 2004

intention to complete Year 12

future intentions (study, apprenticeship, other work etc.).

This analysis also enables us to investigate which characteristics are important factors in determining who works while at school. Regression results for Years 10, 11 and 12 for males and females appear in appendix B (tables 15 to 20).

Not all propensity scores appear in all regressions because in an experimental design context it is impossible to randomise across events, particularly for events that have not yet been observed. As propensity scores are acting as a proxy to the experimental design, it is inappropriate to use the propensity of working in Year 12 on a Year 11 outcome. Instead, we consider the propensity scores

NCVER 13

in the year immediately before the outcome measure. For example, we use the propensity scores for working in Year 10 when considering a Year 11 outcome.

The final stage of our analysis is to undertake a series of regression models to determine the effect of working hours on school and post-school outcomes. The four investigations undertaken are (separately for males and females):

Year 11 and 12 retention: logistic regression of retention to Year 11, and Year 12 against working hours in Year 10 (for Year 11) and working hours in Year 11 (for Year 12)

Ordinary least squares (OLS) regression of Year 12 TER score against working hours in Year 12

Full-time study status in either of the two years post-Year 12 completion: logistic regression against working hours in Year 12

Full-time employment in 2007 (including apprenticeships and traineeships) for those who did not undertake any full-time study in the two years after completing Year 12: logistic regression against working hours in Year 12.

Each regression model uses one or a combination of treatment variables, which categorises the number of hours worked in each school year level between Years 10, 11 and 12.


How many students are working? The Australian Bureau of Statistics (ABS) reports participation in employment amongst 15 to 19-year-olds still at school in its monthly labour force survey. In August9

Figure 1 Proportion of 15 to 19-year-olds at school who are employed, August 1986–2008

2008, for the 800 000 15 to 19-year-olds who were still at school, around a third (or 297 000) of them were working (in either full-time or part-time employment). The proportions combining school and work are illustrated in figure 1.

Source: ABS labour force status (ST LM3) by sex, age (15–24), age 15–19 years only, at school, from April 1986.

Females are more likely to work than males, increasing by around ten percentage points per decade since 1986. Their rates have declined, however, since peaking at 40.5% in 2006. The trend for males is similar to females, although a little more modest, peaking at 30.5% in 2001. The proportion of males combining school and work has remained relatively constant over the last two years.

The Y03 LSAY cohort is asked at the time of their annual interviews, Do you currently work in a job, your own business or on a farm? Details of up to three jobs are recorded, and the main job is identified. They are then asked, for each job, Altogether, how many hours do you usually work each week in your present job? If hours vary, they are asked, In your last four weeks of work, how many hours per week, on average, have you worked including paid holidays? The salary is also recorded. In this report we use the average weekly hours summed across all the jobs a young person works. By combining their self-reported work activity across the survey waves (as this is an aged-based cohort), we are able to assess the proportion of students combining school and work in a given school year level. Since the majority of students left school in 2008, we use 2007 as the cut-off point.

9 August is selected because it coincides with the predominant LSAY survey period.

0

5

10

15

20

25

30

35

40

45

1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008

Year

% e

mpl

oyed

Male

Female

NCVER 15

Table 1 Percentage of respondents working in each school year level, Y03, 2003–07

Working Year 9 Year 10 Year 11 Year 12

Average age at interview 15.7 16.7 17.7 18.7 Males (n) 204 1903 2069 1654

% working 39.3 47.1 51.0 51.9

Average hours worked* 11.8 12.8 12.4 12.1

Females (n) 168 2193 2623 2230

% working 45.3 54.4 60.3 62.4

Average hours worked* 9.9 11.4 11.2 10.8

All (n) 372 4096 4692 3884 % working 41.8 50.7 55.8 57.4

Average hours worked* 10.9 12.1 11.7 11.4 Note: * Based on only those who are in employment while undertaking the given school year level. It also excludes those

whose working status is undefined, or who stated they worked for more than 40 hours a week when at school.

For the Y03 cohort, around half of senior secondary students indicated that they had employment while at school, and this proportion rises with increasing school year level.

As illustrated by comparing figure 1 with table 1, the LSAY data report more work activity by school students than the ABS labour force data. This has also been the case with previous LSAY research. Vickers, Lamb and Hinkley (2003) reported that 25% of school students in Year 9 in 1995 combined school and work, while Robinson (1996) reported that 30.5% of males and 40.3% of females who were aged 17 in 1992 and in Years 11 and 12 combined school and work.

Two explanations come to mind: the first relates to differences in the sample populations, the second to the definition of ‘work’. The focus of this paper is the effect of working while at school on outcomes: while we acknowledge differences in the estimates of the proportions of school students working, these absolute differences are not an important factor for this research.

Distribution of hours worked Figure 2 shows the distributions of the hours worked in each school year level for all students in the LSAY Y03 cohort between 2003 and 2007 (with separate analysis in appendix A for males and females). The box plots describe the distribution of hours worked, with the tails describing the range (smallest and largest) of hours worked, and the box describing the lower quartile, median and upper quartile. The ‘+’ is the mean value. The hours of work in excess of 40 hours have been considered outliers and have not been included in the box plots.


Figure 2 Box plot of working hours for all respondents by school year level, Y03 cohort

As illustrated in figure 2, the distribution of hours worked across the school year levels is virtually the same between Years 9 and 12. Overall, the distribution of hours worked does not appear to vary greatly between school year level, with the mean number of hours worked being between 11 and 12 hours for each of the four school year levels. The students who work in Year 12 work marginally fewer hours than those who work in Year 10.

Table 2 presents the sample sizes and mean hours worked for all respondents in Years 9 to 12, as well as the number of students who are working longer hours (≥ 15 hours per week). Of all those working, up to 20% are working more than 15 hours per week, although there is a slight decline in this percentage for students who are in Year 12. For those who are working more than 15 hours per week, they are working on average up to 20 hours per week.

Table 2 Summary statistics of working and working hours by year level, Y03 cohort

Year 9 Year 10 Year 11 Year 12

No. in year level 890 8077 8405 6762

No. working in year level 372 4096 4692 3884

% of all students working 41.8 50.7 55.8 57.4

Mean working hours (for those working) 10.9 12.1 11.7 11.4

No. working ≥ 15 hours per week 57 817 874 664

Mean working hours (≥ 15 hours per week) 20.2 20.4 20.1 20.7

% of all students working ≥ 15 hours per week 6.4 10.1 10.4 9.8

% of working students working ≥ 15 hours per week 16.7 21.0 19.2 17.7

Note: All figures are unweighted to provide an indication of the absolute level of working and working hours.

NCVER 17

Figure 2 and the box plots by gender in appendix A demonstrate that males on average work longer hours than females, but a higher proportion of female students combine school and work, with 52% of males and 62% of females working in Year 12. The analysis of the Y03 data finds that slightly more young people work when in Year 12 than in Year 11. This is consistent with Robinson (1996), using data from the Youth in Transition surveys,10

10 Prior to the current program, LSAY was based on two other annual surveys; the Australian Youth Survey (AYS,

1989–97), and the Youth in Transition survey (YIT), both of which were age-based cohorts.

who concluded that in the mid-1990s a quarter of students combined school and work in Years 9 and 10, rising to a third of students in Years 11 and 12.


Who combines part-time

work and school? We use previous research to select the characteristics for modelling who is most likely to combine school and work (by school year level), and from these models use the propensity scores to control for background characteristics in the later outcomes modelling. The characteristics of those most likely to combine school and work identified in previous research have generally been consistent, and are summarised in table 3.

Table 3 Summary of characteristics of those who combine school and work from previous research

Study Data Population Proportion working part-time

Characteristics of those who combine school and work

Biddle (2007) Census (characteristics for 2001 census, proportions working provided for 1986, 1991, 1996 & 2001)

15 to18/19-year-old high school students

1986: 10% males, 14% females, 12% overall 1991: 15% males, 22% females, 18% overall

1996: 19% males, 28% females, 24% overall 2001: 23% males, 32% females, 28% overall Two-thirds work less than 10 hours per week

Females > males 17 yr olds < 18 yr olds ACT > Qld > NT > Vic.= WA > NSW >

SA > Tas ESB > NESB Non-Indigenous > Indigenous Govt school > Catholic > independent

school High SES > low SES Income in 3rd and 4th quartile work

longer hours Parents have no degree > parents with

secondary education or higher Couples > single parent families Metro > rural, but longer hours in rural

area ESB > NESB, but those in NESB who

work, work longer hours Longer hours for those whose parents

have no degree and those who live in remote areas

Students in retail (food) work longer hours

Howieson et al. (2006)

10% survey of S3 to S6 students in Scottish state and independent schools

N = 20 700 surveyed between 2003 and 2006

School levels S3 to S6, students aged 15–18, comparable to Y9–Y12 Australian school years

S3 48%, S4 56%, S5 64%, S6 83% 59% overall

Average hrs per week:

S3 7.3%, S4 9.3%, S5 10.7%, S6 12.5% 2/3 worked 1–10 hours per week

Rural > metro Females > males Little difference by SES, but those in

lowest SES < others Those with more certain career plans >

those with no clear idea of career path Disenchantment with school not related

to part-time work More active social life > less active

social life

Vickers et al. (2003)

Y95 LSAY cohort

Y9 in 1995 26.1% males, 23.7% females Average hours of work = 8.6 hours

Males > females ESB > NESB Rural > metropolitan Low SES < other quartiles

Robinson (1999)

Youth in Transition (YIT) survey, precursor to current LSAY, year-based cohorts

Aged 17 in 1992, effect of working in Year 12 in 1994

40% females, 30% males in part-time employment Average 9 hrs per week

Did not report on characteristics, but found that workers generally happier with money they get each week, independence, but not what they can do in their spare time

NCVER 19

Study Data Population Proportion working part-time

Characteristics of those who combine school and work

Robinson (1996)

Youth in Transition (YIT) survey, precursor to current LSAY, year-based cohorts

Years 8–12 from 1989 to 1992

1989: 24.2%

1990: 27.8%

1991: 32.5%

1992: 35.4%

Increases with school year level, but drops Y11 to Y12

Females > males (except in 1989) White collar > semi-skilled and

unskilled Wealthier families > poorer families Parents with secondary education > no

secondary education > parents with degree

Government schools > independent schools

Higher self-perception of academic ability > lower perceived academic ability

Intend to study only post-school > combine post-school study and work

Females, more so than males, tend to combine school and work, as do those in higher rather than lower socioeconomic status quartiles, and those from English speaking backgrounds. Students whose parents are working are more likely to combine work and school, but the types of jobs their parents do can also have an impact. Students with parents in white-collar jobs are more likely than those with semi-skilled or unskilled parents to combine work and school. This could relate to work ethic, the impact of government benefits, as well as the networks of prospective employers their parents can supply. Indeed, the most common way the Y03 LSAY cohort found a job in 2007 (when they were aged 19–20 years) was through a friend or relative.

Apart from gender and socioeconomic status, many of these reported characteristics are also associated with early school leaving (Curtis & McMillan 2008), which makes it difficult to separate out the effects of part-time work on school and post-school outcomes.

Characteristics of students who combine school and work and propensity score regression Based on the findings of previous research (table 3), separate regression models for each gender and school year level were run on the binary response variable, working or not working, against the following background characteristics:11

socioeconomic status

locality

school type

post-school plans

receipt of youth allowance

intention to complete Year 12 asked at age 15

Participation in VET in Schools in 2004

academic ability (scores in maths, problem-solving, science, reading) at age 15.

The propensity scores were calculated from these regressions for each individual, indicating their probability of combining school and work. These were then used to control for background

11 We would also have liked to include an outcome measure of personal attributes and qualities, such as individual

motivation, health or behaviours that could affect outcomes, but were unable to do so because such information is not well measured in LSAY. This could be an area for future research with other longitudinal datasets such as the Australian Temperament Project (ATP) or the Youth in Focus dataset.


characteristics in subsequent statistical modelling in this paper. More details on these regressions are contained in appendix B.

Those who combine school and work tend to have post-school aspirations of apprentice and traineeships, are more likely to be in the second highest SES quartile, tend not to be in receipt of Youth Allowance, and are more likely to attend Catholic (or government for females who work in Year 12) schools. Students who live in remote or regional locations are more likely to work than those living in metropolitan locations.

These results are consistent with other research in this area (refer table 3), with the exception of locality. This difference may relate to our definition of work including ‘work in a job, your own business or on a farm’, which could, depending on the timing of the interview (most LSAY interviews are conducted between July and December each year), include seasonal work, which is more prevalent in regional localities.

Academic ability appears only to be an important predictor for males working in Year 11, and intention to complete Year 12 is not a predictor for males or females combining school and work. This is worth noting here because in many other LSAY research reports, academic ability and intention are strong predictors of school and post-school outcomes (Fullarton 2002; Lamb & McKenzie 2001; Marks, McMillan & Hillman 2001). The large variety of young people combining school and work may partially explain these findings. VET in Schools participation is not associated with an increased propensity to combine school and work.

Post-school intention is a good predictor of likelihood to combine school and work, especially in Year 12. For both males and females, those intending to undertake apprenticeships or traineeships, or those who intend to join the workforce soon after leaving school are more likely to combine school and work. Conversely, those who are intent on post-school study, either at TAFE, university or with some other training provider, are less likely to combine school and work.

Average hours worked by student characteristics Since we see little variation in the characteristics of students by school year level, we provide the average hours worked by characteristics for only those significant characteristics, and only for working in Year 12.

Table 4 highlights that students who combine school and work are a reasonably homogeneous group, in terms of work intensity, with limited variation in hours of work by background characteristics, aside from locality. Year 12 male students living in remote areas work for relatively more hours when in Year 12 (14.0) than their metropolitan counterparts (11.8), but this trend is not evident for females.

However, where there is variation, those most likely to work do not always work the longest hours. For example, we know from figures 1 and 2 that females are more likely to work in Year 12 than males, but males work on average longer hours (12.1) than females (10.8). Similarly, those from a medium-high socioeconomic status are most likely to work, but work on average fewer hours than those from lower socioeconomic status quartiles. Receipt of Youth Allowance does not appear to be a good differentiator of average hours of work in Year 12.

Students with post-school plans that relate more to employment (job, apprenticeship, traineeship) work on average longer hours than Year 12 students with more academic post-school plans (university, TAFE or other training).

Those intent on university work the least number of hours in Year 12. Students with university intentions may be moderating their work to gain better Year 12 results, whereas students who have post-school employment plans may have already begun to be less interested in school, and be intentionally forming a stronger attachment to the labour market.

NCVER 21

Table 4 Average hours worked in Year 12 by student characteristics, by gender

Male Female

Characteristic Mean hrs of work per week Mean hrs of work per week

Locality Metropolitan 11.8 10.7

Regional 13.0 11.0

Remote 14.0 10.4

School sector Government 12.8 11.3

Catholic 11.3 10.4

Independent 10.8 9.1

SES Low SES quartile 13.6 11.1

Low-medium SES quartile 12.4 11.3

Medium-high SES quartile 11.7 10.9

High SES quartile 11.0 10.0

Post-school intentions Go to university 9.9 9.5

Get an apprenticeship 13.5 12.8

Get a traineeship 12.7 10.7

Go to a TAFE college 11.6 11.4

Do some other course or training elsewhere 12.9 9.0

Look for work/get a job 12.4 11.5

Other 15.9 12.3

Don't know 10.2 11.9

Receive Youth Allowance or ABSTUDY No 12.1 10.6

Yes 12.2 11.6

Don't know 12.4 9.9


School outcomes We explore the impact of different hours worked on school retention to Years 11 and 12, and on Year 12 performance, measured as TER score.

Impact on school retention First, we investigate the effect of working on school year level retention between Years 10 and 12. In particular, we model retention to Year 11 against working hours in Year 10 and retention to Year 12 against working hours in Year 11. We have elected to look at retention to Year 12 rather than Year 12 completion, because the majority of students in the LSAY Y03 sample who commence Year 12 go on to complete it.

The Y03 LSAY cohort has a male Year 11 to Year 12 retention rate of 85% and a slightly higher female retention rate of 88%,12

By modelling retention to these two separate school year levels rather than as a single measure from Year 10 to Year 12, we are able to assess the impact of combining school and work at two separate decision points in the school-to-work transition.

which provides some variation with which to model the effect of working hours on retention.

The results (predicted probability of continuing to Year 11 from Year 10) of the logistic regressions of hours worked in Year 10 on Year 11 retention are presented in table. The predicted probabilities for retention are calculated for each of the categorical classification of hours worked by applying the regression model values separately for males and females.13

Table 5 Predicted probability of Year 10 to Year 11 retention by hours worked in Year 10

More details of the regression models are contained in appendix B, tables 21–24.

Working hours Males Diff. from 0 Females Diff. from 0

Year 10 Not working 0.83 - 0.85 -

0 < x < 5 0.84 +0.01 0.88 +0.03

5 = 20 0.59* -0.24 0.70* -0.15

Note: * significantly different from not working.

12 The Y03 cohort has a Year 12 completion rate of 83%, which is significantly higher than the national average reported

by the ABS of around 75% (ABS 2008). 13 The predicted probabilities are calculated, based on the results of the regression at the average propensity score and for

each level of working hours with the other hours set to zero.

NCVER 23

For males, working more than five hours while in Year 10 leads to a lower Year 11 retention rate of between -6 and -24 percentage points, whereas females can work up to 15 hours before the negative effects are observed, and with lesser impact (between -7 and -15 percentage points).

Turning to retention to Year 12, we see a lesser effect for males than we did with Year 11 retention, with hours worked in Year 11 not affecting Year 12 retention (not statistically significant) until the hours exceed 20 hours a week, and here the penalty is of the order of -8 percentage points.

Table 6 Predicted probability of Year 11 to Year 12 retention by hours worked in Year 11

Working hours Males Diff. from 0 Females Diff. from 0

Year 11 Not working 0.86 - 0.88 -

0 < x < 5 0.85 -0.01 0.88 0.0

5 < = x < 10 0.85 -0.01 0.92 +0.04

10 < = x < 15 0.84 -0.02 0.88 0.0

15 < = x < 20 0.84 -0.02 0.86* -0.02

X > = 20 0.78* -0.08 0.75* -0.13

Note: * significantly different from not working.

For females, working more than 15 hours in Year 11 increases the probability of leaving school prior to undertaking Year 12 by a couple of percentage points for 15–20 hours, and by 13 percentage points for more than 20 hours of work. Again, as for males, the effect of combining work and study is not as strong for retention to Year 12 as it is for retention to Year 11.

Impact on school performance (TER score) Ordinary least squares (OLS) regressions were used to investigate the effect of hours of work in Year 12 on Year 12 performance measured using TER scores. These regressions considered only those in Year 12 who actually obtained a TER score.14

The results of the regressions, presented as adjusted mean TER are contained in table 7 for males and table 8 for females. (Full results are presented in appendix B.)

The interest in this section lies with students who are choosing an academic pathway.

Table 7 Mean TER scores by hours worked in Year 12, males

Hours worked Mean TER Difference from not working

95% confidence Interval

Not working 75.5 - (74.5, 76.5)

0 < x < 5 75.4 -0.1 (72.5, 78.1)

5 < = x < 10 73.4* -2.1 (71.4, 75.3)

10 < = x < 15 72.1* -3.4 (70.0, 74.1)

15 < = x < 20 73.8^ -1.7 (70.1, 76.9)

X > = 20 70.0* -5.5 (67.1, 73.0) Notes: * significantly different from not working.

^ the lack of statistical significance is due to sample size and variation in TER scores.

Working a small number of hours (fewer than five) has no detrimental effect on Year 12 achievement for males, but working longer than five hours can reduce a respondent’s TER score.

14 It is possible for a student to complete Year 12 and not obtain a TER; however, we restrict the analysis to those with a

TER score because the focus of our analysis is on the impact of combining school and work on TER score.


The difference between not working, and working for more than 20 hours a week for males is on average a reduction of -5.5 TER points.

Table 8 highlights that females can work up to ten hours a week in Year 12 before it affects Year 12 performance, but once this threshold is exceeded, the TER performance falls significantly. Females appear to be better able to manage the competing demands of Year 12 and working up to ten hours a week, with their TER scores affected at higher working hours than males. (The effect of working more than 20 hours a week has a similar effect to males, reducing female TER scores on average by 4.4 points.)

For both males and females, the detrimental effect of working on TER scores is not linear. Working between 15 and 20 hours appears to have a lower impact on TER scores than working between 10 and 15 hours for males, and for females, working more than 20 hours a week has a lower impact on TER scores than working between 15 and 20 hours a week. In both cases, these scores are not statistically significant and remain lower compared with TER scores for those not working at all. Therefore, we can conclude that there is a generally negative impact on TER scores associated with working for longer than five hours a week in Year 12.

Table 8 Mean TER scores by hours worked in Year 12, females

Hours worked Mean TER Difference from not working

95% confidence Interval

Not working 78.1 - (77.1, 79.1)

0 < x < 5 78.5 +0.4 (76.3, 80.7)

5 < = x < 10 77.9 -0.2 (76.6, 79.3)

10

NCVER 25

Post-school outcomes In this section we look at the effect of working in Year 12 on post-school outcomes for students who have completed Year 12 but who have gone on to post-school full-time study or post-school full-time employment in 2007, that is, one to two years after Year 12 for the majority of students.15

We restrict our analysis to only Year 12 completers so as not to contaminate the analysis with early school leavers, as time in the labour market matters when assessing employment outcomes. A separate analysis of the effects of combining school and work on early school leavers’ post-school outcomes is perhaps an area for future research.

Effect of working in Year 12 on post-school full-time study status The outcome of interest is whether or not a respondent undertook any post-school full-time education in the two years after completing Year 12. Separate logistic regressions were undertaken for males and females, and the regressions consider the effect of TER score because we know this has an impact on post-school study. Full results appear in appendix B.

The probabilities of being in full-time study by hours of work in Year 12 are presented in tables 9 and 10. The probabilities are calculated at the averages of the TER and propensity scores.

Table 9 Predicted probability of undertaking full-time post-school study for hours worked in Year 12, males

Working hours Pr (full-time study) Difference from not working

Not working 0.68 -

0 < x < 5 0.66 -0.02

5 < = x < 10 0.67* -0.01

10 < = x < 15 0.59 -0.09

15 < = x < 20 0.58 -0.10

X > = 20 0.52 -0.16

Notes: * significantly different from not working. the statistical significance is influenced by sample size. There are a greater number of respondents in the 5–10 category and, therefore, the observable significant difference can be smaller. The overall trend is what is interesting in this table.

The effects of working in Year 12 for males (table 9) and females (table 10 ) on post-school full-time study are very different.

For males, the general trend was that the more hours worked in Year 12, the less likely they were to undertake post-school full-time study. Working for more than 20 hours in Year 12 reduced the probability that a male would pursue full-time post-school study by -16 percentage points.

15 The effect of combining school and work in Year 12 is only considered as this is the year most immediate to the post-

school outcome.


Table 10 Predicted probability of undertaking full-time post-school study for hours worked in Year 12, females

Working hours Pr (full-time study) Difference from not working

Not working 0.66 -

0 < x < 5 0.82* +0.16

5 < = x < 10 0.77* +0.11

10 < = x < 15 0.70* +0.04

15 < = x < 20 0.53 -0.13

X > = 20 0.65* -0.01 Note: * significantly different from not working.

For females, we find that, unlike for males, working for a moderate number of hours (less than 15 hours a week) in Year 12 can have a positive impact on the probability that they will go on to pursue post-school full-time study. However, once hours exceed 15–20 hours a week, then as for males, a negative effect is evident.

While we do not know the reason for this, females may be better able to manage the conflicting demands of school and work (as also evidenced in TER results, where females can work slightly longer hours) than their male counterparts.

Effect of working in Year 12 on post-school full-time employment status Finally, we investigated the impact of hours worked during Year 12 on full-time employment status in 2007 for those who completed Year 12, but who did not undertake any full-time study in the two years after completing Year 12. As this is an age-based cohort, the majority of students had one to two years in the labour market by 2007.

Logistic regressions were undertaken for the dichotomous variable, in full-time employment or not in full-time employment in 2007 (results appear in appendix B, table 37–40).

Tables 11 and 12 present the predicted probabilities of being in full-time employment in 2007 for males and females separately by hours worked in Year 12.

Table 11 Predicted probability of full-time employment with no post-school study for hours worked in Year 12, males

Working hours Pr (full-time employment) Difference from not working

Not working 0.32 -

0 < x < 5 0.46 0.14

5 < = x < 10 0.52* 0.20

10 < = x < 15 0.59* 0.27

15 < = x < 20 0.56* 0.24

X > = 20 0.52* 0.20 Note: * significantly different from not working.

For males who complete school and pursue no post-school study, working for more than five hours in Year 12 is beneficial over not working at all. However, the rates of return do not increase in a linear manner, and working between 10 and 15 hours a week maximises the probability of better post-school employment outcomes.

NCVER 27

Table 12 Predicted probability of full-time employment with no post-school study for hours worked in Year 12, females

Working hours Pr (full-time employment) Difference from not working

Not working 0.20 -

0 < x < 5 0.26 0.06

5 < = x < 10 0.32* 0.12

10 < = x < 15 0.39* 0.19

15 < = x < 20 0.49* 0.29

X > = 20 0.38* 0.18 Note: * significantly different from not working.

For females, we see a similar pattern with positive benefits of combining school and work in Year 12 on post-school employment outcomes. However, females have to work for slightly longer hours (15 to 20 hours a week in Year 12) to gain maximum benefit (of +29 percentage points), whereas maximum benefits are realised for males who work between 10 and 15 hours a week (of +27 percentage points).

Summary These results point to a slightly negative effect of combining school and work on post-school full-time study, apart from a rather unexplained positive effect for moderate hours of work in Year 12 for females. However, once hours of work exceed 15 hours a week, we see a negative effect, as for all hours of work for males. The magnitude of the effects appears to be slightly greater for males than females.

Unlike the negative effects we see for school and post-school study outcomes, we see positive effects from working in Year 12 on post-school employment for both males and females who do not go on to post-school full-time study. The magnitude of these positive effects is consistent for males and females.


Discussion The research in this paper confirms the findings of other research, that students who combine school and work are spread right across the school population, although some groups have a tendency to work longer hours than others. With such a large proportion of students combining school and work, it is not surprising that they do not have a set of strong defining characteristics. However, we do find that students who combine school and work are in general in the higher, but not highest socioeconomic status quartile, attend Catholic or government schools, are not in receipt of Youth Allowance, and have a preference for an apprenticeship, traineeship or a job when they leave school.

This paper finds that the effects of combining school and work of more than ten hours a week are moderately negative on school and post-school study outcomes, but positive on post-school full-time employment. These findings are similar to the earlier work of Robinson (1999), but what is interesting is that the two studies use different cohorts of young people in quite different economic conditions. The earlier research focused on young people in the Youth in Transition survey in a period of economic downturn (aged 17 years in 1992), whereas the analysis in this paper concerns a group of young people from the LSAY Y03 cohort who were aged 15–19 years between 2003 and 2007, in a much stronger economic climate. Despite the differences in economic conditions, the different cohorts and the growth in the numbers of young people combining school and work (increasing from around a quarter of 15 to 19-year-olds in 1992 to around a third in 2008), we find the same effects for combining school and work on school and post-school outcomes.

The novelty of the approach in this paper is the way in which school outcomes are measured. In addition to modelling Year 12 completion by gender and hours worked (as in Robinson 1999), we decompose it into retention from Year 10 to Year 11 and then retention from Year 11 to Year 12. This enables us to better understand the way combining school and work can affect the decision points between Years 10 and 12, while allowing us to more finely model the effect of work in previous school year levels. This approach uncovered the finding that the negative effects of combining school and work on school retention are stronger for those who work in Year 10 than those who work in Year 11. Perhaps this is because those who are working in Year 11 tend to moderate their hours. But, overall, the negative effects of combining school and work are modest, unless the person is working excessive hours (over 15–20 hours a week).

Is combining school and work detrimental to school and post-school outcomes? The results in this paper point towards moderate hours being preferable. Longer hours appear to be detrimental for educational outcomes but good for employment outcomes, which tends to suggest that those willing to work the longer hours are distancing themselves from the education environment.

NCVER 29

References ABS (Australian Bureau of Statistics) 2008, Schools Australia 2007, cat.no.4221.0, ABS, Canberra, December

2008, . ——2008, Labour force Australia, cat.no.6202.0, ABS, Canberra, viewed December 2008,

. Biddle, N 2007, ‘The labour market status of Australian students: who is unemployed, who is working and for

how many hours?’, Journal of Education and Work, vol.20, no.3, July 2007, pp.179–209. Billett, S 2006, Informing post-school pathways: investigating school students’ authentic work experiences, NCVER,

Adelaide. Curtis, D & McMillan, J 2008, School non-completers: profiles and destinations, LSAY research report 54, ACER,

Melbourne. Fullarton, S 2002, Student engagement with school: individual and school-level influences, LSAY research report 27,

ACER, Melbourne. Howieson, C, McKechnie, J & Semple, S 2006, The nature and implications of the part-time employment of secondary

school public, Scottish Executive Social Research, Department of Enterprise, Transport and Lifelong Learning, Scotland.

Lamb, S & McKenzie, P 2001, Patterns of success and failure in the transition from school to work in Australia, LSAY research report 18, ACER, Melbourne.

Marks, GN, McMillan, J & Hillman, K 2001, Tertiary entry performance: the role of student background and school factors, LSAY research report 22, ACER, Melbourne.

Marsh, HW, & Kleitman, S 2005, ‘Consequences of employment during high school: character building, subversion of academic goals, or a threshold’, American Educational Research Journal, vol.42, issue 2, p.331.

Meyerhoff, MK 2006, ‘Part-time work for teens’, Pediatrics for Parents, vol.22, issue 10, pp.8–10. Rosenbaum, P & Ruben, D 1983, ‘The central role of the propensity score in observational studies for causal

effects’, Biometrika, issue 70, pp.41–55. Robinson, L 1996, School students and part-time work, LSAY research report 2, ACER, Melbourne. ——1999, The effects of part-time work on school students, LSAY research report 9, ACER, Melbourne. Shanahan, MJ & Flaherty, BR 2001, ‘Dynamic patterns of time use in adolescence’, Child Development, 72,

pp.385–401. Smith, E & Green, A 2005, How workplace experiences while at school affect career pathways, NCVER, Adelaide. Staff, J & Mortimer, JT 2007, ‘Educational and work strategies from adolescence to early adulthood:

consequences for educational attainment’, Social Forces, vol.85, issue 3, pp.1169–95, Chapel Hill. ——2008, ‘Social class background and the school to work transition’, New Directions for Child and Adolescent

Development, no.119, pp.55–69. Vickers, M, Lamb, S & Hinkley, J 2003, Student workers in high school and beyond: the effects of part-time employment on

participation in education, training and work, LSAY research report 30, ACER, Melbourne.

http://www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/4221.02007?OpenDocument�http://www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/6202.0Feb%202009?OpenDocument�


Appendix A Data This research uses data from the LSAY Y03 cohort. The Y03 cohort follows 10 370 students from 2003, when they were 15 years of age. The pathways of these young people as they move through their senior secondary school years into post-school education and post-school employment are surveyed. Data from the period 2003 to 2007 are used in this paper.

Because the Y03 cohort is an aged-based, rather than a cohort based on school year level, we have a spread of students at different school year levels in any given calendar year. This is important, because in our analysis we are focusing on working in different school year levels, so we must sum this activity across the years of interview (waves). The table below summarises the LSAY Y03 cohort by calendar year and school year level up to the most recently available survey wave (2007 interviews).

Table 13 LSAY Y03 data by school year level and year of data collection (weighted)

School level 2003 2004 2005 2006 2007

Avg. age 15.7 Yrs 16.7 Yrs 17.7 Yrs 18.7 Yrs 19.7 Yrs n % n % n % n % n %

Year 9 901 8.7 11 0.1 0 0.0 0 0.0 0 0.0

Year 10 7 451 71.9 714 7.6 8 0.1 0 0.0 0 0.0

Year 11 1 979 19.1 5 769 61.5 611 7.0 9 0.1 0 0.0

Year 12 39 0.4 1 628 17.4 4 940 56.8 436 5.7 6 0.1

Left school 0 0.0 1 257 13.4 3 131 36.0 7 275 94.2 6 652 99.9

Total 10 370 100.1 9 379 100.0 8 690 99.9 7 720 100.0 6 658 100.0 Note: * totals do not always sum to 100 due to rounding.

Definition and scope The definition of combining school and work is derived from the LSAY respondent’s answer to the following question, which asks at the time of the survey:16

The population of interest for the analysis is all students in the Y03 cohort for the current waves (2003 to 2007), where we can establish if they did or did not work in the school year level of interest. When looking at school outcomes, we consider all of these students. For the analysis on TER score, we include only those students who reported a TER score. For the analysis on post-

Do you currently work in a job, your own business or on a farm? This is combined with the questions on whether or not they are at school, and in which school year they are in, to derive variables for combining school and work across the different school years. For young people with more than one job, the hours worked are the sum of all hours worked per week across all jobs (at the time of the survey).

16 LSAY interviewing is conducted from July/August – December/January each year, and so this will include school

holiday jobs for some young people.

NCVER 31

school outcomes, we consider all students who, two years after completing Year 12, go on to either full-time post-school study or full-time employment.

Distribution of hours of work by school year level Figure 3 shows that there is very little change in the working hours of both girls and boys over the four school year levels. On average, both girls and boys have median working hours of around ten hours per week. However, it appears as though there are more males who are working longer hours. From table 14, we see that there are is a higher percentage of males working more than 15 hours per week, particularly for those who worked in Year 9 and Year 12, with 21% of males and 14% of females working more than 15 hours in Year 12. The mean number of hours worked by students working long hours is around 20 hours per week, with very little difference between males and females.

Figure 3 Box plot for working hours by school year level by gender, Y03 cohort

Table 14 Summary statistics of working hours and numbers by year level by gender

Male Female

Year 9 Year 10 Year 11 Year 12 Year 9 Year 10 Year 11 Year 12

No. in year level 519 4043 4055 3186 371 4034 4350 3576

No. working 204 1903 2069 1654 168 2193 2623 2230

% of all students working 39.3 47.1 51.0 51.9 45.3 54.4 60.3 62.4

Mean hours worked 11.8 12.8 12.4 12.1 9.9 11.4 11.2 10.8

No. working ≥ 15 hours per week 37 417 449 351 20 400 425 313

% of all students 7.1 10.3 11.1 11.0 5.3 9.9 9.8 8.8

% of working students 18.1 21.9 21.7 21.2 11.9 18.2 16.2 14.0

Mean hours (≥ 15 hours) 22.0 21.5 21.2 21.4 17.5 19.3 19.0 20.1


Appendix B This appendix contains the results of the regression models. All statistical analysis is carried out using the SAS statistical package.

The following summarises the definition of each output measure for logistic regression:

• b: These are the estimated beta coefficients for the logistic regression equation for predicting the dependent variable from the independent variables. The logistic prediction equation is

)exp1/(1 zp −+=

Where nn xbxbbz ×++×+= 110

• SE: The standard errors of the regression coefficients

• Wald and Sig.: Provide the Wald Chi-Square Statistic ((coefficient/S.E)2) and the two-tailed p-value used in testing to determine whether the coefficient is significantly different from 0 (the reference category).

• df: This column lists the degrees of freedom for testing the coefficients.

• Odds ratio: These are the odds ratios for predictors. They are simply the exponentiation of the coefficients. Odds ratios of greater than one indicate a higher chance of the event occurring than the reference group; odds ratios of less than one indicate a lower chance than the reference group. The confidence interval for odds may also be presented; if this confidence interval contains one, then we can conclude that this effect has the same influence on the response as the reference category.

Tables 13 to 18 contain the results of the logistic regression, which model the probability of working in Years 10, 11 and 12 by gender. Separate models are run for each school year level. The propensity scores for these regressions are then used to summarise the background information of respondents into a single value. These provide a method for reducing selection bias in the modelling of our treatment effects of hours worked. The propensity scores are calculated as the probability that an individual will work, given the known background characteristics, that is, they ‘average’ out the effects of the background characteristics. These propensity scores are included as covariates in the subsequent regression models (tables 23 to 42) used in the school and post-school outcomes. Propensity score regression assesses the importance of intensity of work after removing the background effects.

Not all propensity scores are used in all regressions. For example, it is not appropriate to include the propensity to work in Year 12 when investigating retention to Year 12. In this case, you would only include propensity to work in Years 10, and 11. Note that propensity score regression coefficients are not examined for significance in the final regression analysis.

NCVER 33

Table 15 Regression results for working in Year 10: males, Y03, 2003–07

Characteristic Coefficients (response reference

category is working)

S.E Wald df p-value Odds ratio

95% CI for odds ratio

Maths achievement quartile* Lowest -0.200 0.180 1.230 1 0.268 0.819 (0.576, 1.166)

Second 0.026 0.144 0.032 1 0.859 1.026 (0.774, 1.361)

Third 0.176 0.114 2.393 1 0.122 1.192 (0.954, 1.490)

Highest Reference category

Problem-solving achievement quartile Lowest 0.079 0.179 0.193 1 0.660 1.082 (0.762, 1.537)

Second 0.113 0.144 0.616 1 0.433 1.120 (0.844, 1.484)

Third -0.010 0.114 0.007 1 0.932 0.990 (0.792, 1.239)


Science achievement quartile Lowest -0.177 0.169 1.104 1 0.293 0.838 (0.602, 1.166)

Second -0.125 0.140 0.793 1 0.373 0.882 (0.670, 1.162)

Third -0.073 0.115 0.409 1 0.523 0.929 (0.742, 1.163)


Reading achievement quartile Lowest 0.215 0.166 1.661 1 0.198 1.239 (0.894, 1.717)

Second 0.076 0.139 0.296 1 0.586 1.078 (0.822, 1.416)

Third 0.107 0.116 0.847 1 0.357 1.113 (0.886, 1.397)


Location* Metropolitan -0.744 0.232 10.239 1 0.001 0.475 (0.301, 0.750)

Regional -0.456 0.237 3.691 1 0.055 0.634 (0.398, 1.009)

Remote Reference category

Sector* Government 0.172 0.099 3.064 1 0.001 1.188 (0.980, 1.441)

Catholic 0.446 0.419 1.136 1 0.287 1.452 (1.172, 1.799)

Independent Reference category

Socioeconomic status (ISCED, father’s or mother’s if missing)* Low SES quartile -0.063 0.095 0.442 1 0.506 0.939 (0.780, 1.130)

Low-medium SES quartile

-0.249 0.096 6.817 1 0.009 0.779 (0.646, 0.940)

Medium-high SES quartile

Reference category

High SES quartile -0.331 0.094 12.385 1 0.000 0.718 (0.598, 0.864)

VET in Schools in 2004 No -0.093 0.074 1.562 1 0.211 0.912 (0.788, 1.054)

Yes Reference category

Unknown 0.017 0.170 0.010 1 0.921 1.017 (0.728, 1.421)






Post-school intentions* Go to university -0.135 0.305 0.195 1 0.659 0.874 (0.481, 1.589)

Get an apprenticeship

-0.051 0.316 0.026 1 0.871 0.950 (0.512, 1.764)

Get a traineeship 0.810 0.410 3.894 1 0.049 2.247 (1.005, 5.021)

Go to a TAFE college

-0.248 0.323 0.588 1 0.443 0.781 (0.415, 1.470)

Do some other course or training elsewhere

-0.121 0.396 0.093 1 0.760 0.886 (0.408, 1.926)

Look for work/ get a job

0.044 0.319 0.019 1 0.889 1.045 (0.560, 1.952)

Other 0.194 0.477 0.166 1 0.684 1.215 (0.477, 3.096)

Don't know Reference category

Intention to commence Year 12 No 0.300 0.143 4.399 1 0.036 1.350 (1.020, 1.788)


Don't know -0.055 0.175 0.010 1 0.752 0.946 (0.672, 1.333)

Receive Youth Allowance or ABSTUDY* No 0.144 0.142 1.023 1 0.312 1.154 (0.874,1.525)

Yes -0.410 0.156 6.873 1 0.009 0.664 (0.489, 0.902)


Note: *significant at the 5% level; ISCED = International Standard Classification of Education.

NCVER 35

Table 16 Regression results for working in Year 10: females, Y03, 2003–07





Maths achievement quartile Lowest -0.021 0.176 0.014 1 0.905 0.979 (0.694, 1.381)

Second 0.155 0.144 1.158 1 0.282 1.168 (0.880, 1.549)

Third 0.132 0.118 1.266 1 0.261 1.142 (0.906, 1.438)


Problem-solving achievement quartile Lowest -0.220 0.179 1.511 1 0.219 0.803 (0.566, 1.139)

Second -0.164 0.145 1.273 1 0.259 0.849 (0.639, 1.128)

Third -0.068 0.118 0.337 1 0.562 0.934 (0.741, 1.176)



Second -0.089 0.139 0.404 1 0.525 0.915 (0.697, 1.203)

Third -0.099 0.111 0.791 1 0.374 0.906 (0.729, 1.126)



Second 0.049 0.131 0.142 1 0.707 1.051 (0.812, 1.358)

Third 0.139 0.106 1.725 1 0.189 1.149 (0.934, 1.415)



Regional -0.344 0.216 2.531 1 0.112 0.709 (0.464, 1.083)


Sector* Government 0.377 0.095 15.788 1






Post-school intentions Go to university -0.163 0.303 0.289 1 0.591 0.850 (0.469, 1.539)


-0.292 0.387 0.571 1 0.450 0.747 (0.350, 1.594)

Get a traineeship -0.038 0.381 0.010 1 0.921 0.963 (0.456, 2.032)


-0.236 0.314 0.563 1 0.453 0.790 (0.427, 1.462)


0.285 0.451 0.399 1 0.528 1.330 (0.549, 3.218)


-0.171 0.317 0.290 1 0.590 0.843 (0.453, 1.570)

Other 0.283 0.478 0.352 1 0.553 1.327 (0.521, 3.385)




Don't know 0.202 0.236 0.731 1 0.393 1.223 (0.771, 1.942)

Receive Youth Allowance or ABSTUDY* No 0.233 0.148 2.487 1 0.115 1.262 (0.945, 1.685)

Yes -0.123 0.159 0.603 1 0.438 0.884 (0.648, 1.207)


Note: * Significant at the 5% level.

NCVER 37






Maths achievement quartile* Lowest -0.172 0.237 0.528 1 0.467 0.842 (0.529, 1.339)

Second 0.018 0.180 0.010 1 0.919 1.018 (0.716, 1.448)

Third 0.141 0.135 1.091 1 0.296 1.151 (0.884, 1.499)


Problem-solving achievement quartile Lowest 0.192 0.235 0.667 1 0.414 1.212 (0.764, 1.921)

Second 0.060 0.175 0.117 1 0.732 1.062 (0.753, 1.497)

Third -0.016 0.134 0.014 1 0.905 0.984 (0.757, 1.280)



Second -0.134 0.172 0.605 1 0.437 0.875 (0.625, 1.225)

Third -0.064 0.133 0.229 1 0.633 0.938 (0.722, 1.219)


Reading achievement quartile Lowest -0.100 0.209 0.231 1 0.631 0.905 (0.601, 1.362)

Second 0.084 0.163 0.267 1 0.605 1.088 (0.790, 1.498)

Third 0.117 0.132 0.777 1 0.378 1.124 (0.867, 1.456)



Regional -0.260 0.324 0.640 1 0.424 0.771 (0.409, 1.457)



Catholic 0.527 0.129 16.547 1








-0.180 0.335 0.289 1 0.591 0.835 (0.433, 1.610)



-0.450 0.340 1.748 1 0.186 0.638 (0.328, 1.242)


-0.416 0.415 1.007 1 0.316 0.660 (0.292, 1.487)


-0.169 0.337 0.252 1 0.615 0.844 (0.436, 1.634)

Other 0.095 0.501 0.036 1 0.849 1.100 (0.412, 2.933)


Intention to commence Year 12 No -0.161 0.421 0.146 1 0.703 0.852 (0.373, 1.943)


Don't know 0.092 0.407 0.051 1 0.821 1.096 (0.494, 2.432)


Yes -0.460 0.187 6.063 1 0.014 0.631 (0.438, 0.910)


Note: * Significant at the 5% level.

NCVER 39







Second 0.174 0.166 1.107 1 0.293 1.190 (0.860, 1.647)

Third 0.204 0.131 2.439 1 0.118 1.227 (0.949, 1.586)



Second 0.030 0.169 0.032 1 0.858 1.031 (0.740, 1.437)

Third -0.074 0.133 0.310 1 0.578 0.929 (0.717, 1.204)


Science achievement quartile Lowest 0.076 0.213 0.128 1 0.721 1.079 (0.711, 1.636)

Second 0.022 0.162 0.019 1 0.890 1.023 (0.744, 1.405)

Third 0.109 0.124 0.763 1 0.383 1.115 (0.874, 1.422)


Reading achievement quartile Lowest -0.277 0.214 1.676 1 0.196 0.758 (0.498, 1.153)

Second 0.050 0.160 0.099 1 0.753 1.051 (0.769, 1.438)

Third 0.004 0.123 0.001 1 0.974 1.004 (0.790, 1.276)



Regional -0.344 0.294 1.366 1 0.243 0.709 (0.398, 1.262)



Catholic 0.628 0.131 22.929 1






Post-school intentions Go to university 0.324 0.303 1.151 1 0.283 1.383 (0.765, 2.503)


0.795 0.405 3.848 1 0.050 2.213 (1.001, 4.895)



0.428 0.315 1.852 1 0.174 1.535 (0.828, 2.845)


0.145 0.439 0.109 1 0.741 1.156 (0.489, 2.730)


0.612 0.318 3.694 1 0.055 1.844 (0.988, 3.441)

Other 0.810 0.506 2.558 1 0.110 2.247 (0.833, 6.060)




Don't know -0.143 0.482 0.088 1 0.766 0.867 (0.337, 2.227)


Yes -0.244 0.181 1.813 1 0.178 0.784 (0.549, 1.118)


Note: * Significant at the 5% level

NCVER 41







Second 0.330 0.180 3.360 1 0.067 1.391 (0.977, 1.980)

Third 0.107 0.134 0.636 1 0.425 1.113 (0.856, 1.447)



Second -0.140 0.176 0.631 1 0.427 0.870 (0.616, 1.227)

Third -0.144 0.134 1.157 1 0.2822 0.866 (0.667, 1.125)



Second -0.100 0.172 0.332 1 0.564 0.905 (0.645, 1.270)

Third 0.038 0.133 0.079 1 0.779 1.038 (0.799, 1.348)



Second 0.112 0.164 0.465 1 0.500 1.118 (0.811, 1.543)

Third 0.109 0.133 0.680 1 0.410 1.116 (0.860, 1.447)



Regional 0.294 0.314 0.877 1 0.349 1.342 (0.725, 2.483)


Sector* Government 0.4862 0.119 16.687 1








0.2860 0.348 0.677 1 0.411 1.331 (0.673, 2.631)



-0.0201 0.351 0.033 1 0.954 0.980 (0.493, 1.948)


-0.0922 0.426 0.047 1 0.829 0.912 (0.396, 2.102)


-0.0314 0.348 0.008 1 0.928 0.969 (0.490, 1.916)

Other 0.0487 0.525 0.009 1 0.926 1.050 (0.375, 2.938)


Intention to commence Year 12 No 0.0885 0.450 0.039 1 0.844 1.093 (0.453,2.637)


Don't know -0.2384 0.460 0.269 1 0.604 0.788 (0.320, 1.939)


Yes -0.3674 0.193 3.622 1 0.057 0.693 (0.474, 1.011)


Note: * Significant at the 5% level

NCVER 43






Maths achievement quartile Lowest 0.001 0.219 0.000 1 0.998 1.001 (0.652, 1.536)

Second 0.200 0.169 1.403 1 0.236 1.221 (0.878, 1.699)

Third 0.086 0.131 0.434 1 0.510 1.090 (0.843, 1.409)



Second -0.030 0.172 0.030 1 0.863 0.971 (0.693, 1.359)

Third -0.152 0.133 1.311 1 0.252 0.859 (0.661, 1.115)



Second -0.050 0.164 0.091 1 0.763 0.952 (0.690, 1.313)

Third 0.033 0.125 0.069 1 0.793 1.033 (0.809, 1.319)



Secon

Documents

LSAY work and school - ERIC · 2013. 8. 2. · 29 ANOVA for TER scores against Year 12 working hours, males 47 30 Regression means for TER against Year 12 working hours, ... Chang